The Right Budgeting Approach for Growth-Stage Companies
The first question is which budgeting method to use. The three main approaches have meaningfully different characteristics, and the right choice depends on the stage and complexity of the business.
Incremental budgeting starts from the prior year actuals and applies percentage increases or decreases to each line. It is fast and intuitive, but it embeds the structural assumptions of the prior year into the new budget without scrutiny. For a growth company where the business model, team size, and cost structure may have changed substantially, incremental budgeting risks perpetuating poor spending decisions and failing to account for new cost categories. It is appropriate only for very stable, mature cost lines.
Zero-based budgeting starts from zero for every cost line and requires each item of expenditure to be justified from scratch. It is rigorous and effective at identifying wasteful spending, but it is also extremely time-consuming and is often resisted by department heads. Full zero-based budgeting is typically only viable for large organisations with dedicated finance business partners in every function. At growth stage, a modified version, applying zero-based principles to the largest or most variable cost lines while using incremental for stable items, is more practical.
Driver-based budgeting builds the budget from the key operational drivers of the business: headcount drives people costs, transaction volume drives processing fees, customer count drives support costs. This is the most appropriate approach for growth-stage companies because it makes the assumptions explicit, allows rapid scenario modelling by changing the drivers, and creates a direct link between operational plans and financial outcomes. A well-built driver-based model will automatically reforecast when the operational plan changes, reducing the manual rework associated with every forecast update.
Rolling Forecasts vs Annual Budgets
The Gartner FP&A Survey 2025 found that 67 per cent of high-performing finance functions use a rolling forecast alongside or instead of a traditional annual budget. The case for the rolling forecast is strong: an annual budget set in December becomes progressively less useful as the year advances and conditions diverge from the assumptions.
The most practical approach for growth-stage companies is to run both in parallel: a formal annual budget approved by the board in January, used as the governance baseline for the year, and a quarterly rolling forecast (a 4+8 model: 4 months actual plus 8 months forecast) that is updated each quarter and used for operational decision-making. The rolling forecast is not a budget revision: it does not require board approval and does not change the performance targets against which the management team is measured. It is the current best estimate of where the year will land, used by the CFO and CEO to make real-time decisions.
Building Flexibility: Three Scenarios with Stated Assumptions
A 2026 budget that presents only a single base case is not fit for purpose in the current environment. The CFO should develop three explicitly distinct scenarios, each with clearly stated macro and operational assumptions, so that the board can make decisions contingent on which scenario is materialising.
Base case is the CFO's central estimate of how the year will develop given current information. For 2026, reasonable base case assumptions include: UK GDP growth of approximately 1.2 to 1.5 per cent, Bank of England base rate declining gradually to approximately 4.0 per cent by year end, inflation remaining in the 2.5 to 3.0 per cent range, and no material regulatory changes beyond those already announced. Revenue growth assumptions should be built from the bottom up from the pipeline and account base, not top-down from a growth rate percentage.
Upside case identifies two or three specific positive developments that could occur and models their impact. For most growth companies, the upside case is not a general "everything goes well" scenario but a specific "we close the three named enterprise deals in the pipeline in Q1 and the US expansion ramps faster than expected" scenario. Specificity makes it actionable and credible.
Downside case should model the materialisation of identified risks: a major customer churning, the fundraising round closing 6 months later than planned (which is a credible assumption in a market where Series B and C rounds are taking longer to close than in 2021 to 2022), or a regulatory development that increases compliance costs. The downside case should show the runway implications and identify which cost levers would be pulled.
"The value of scenario planning is not prediction. It is that having thought through the decisions in advance, the CFO can act quickly and decisively when the downside materialises, rather than spending three weeks in reactive mode while cash burns."
Headcount Budgets and AI Productivity
Headcount is typically 60 to 75 per cent of the total cost base at growth stage. Getting the headcount budget right is therefore the single most important modelling task in the whole budgeting process.
The October 2024 National Insurance changes significantly increased the employer cost of each employee. The employer NIC rate increased from 13.8 per cent to 15 per cent, and the threshold at which employers begin paying NIC fell from £9,100 to £5,000 per annum. For a company with 40 employees at an average salary of £60,000, this increase adds approximately £30,000 to £40,000 in annual NIC cost. It must be explicitly included in 2026 headcount budgets, particularly for companies that set their 2025 budget before October 2024.
The more complex headcount question for 2026 is the impact of AI on team productivity. The Gartner FP&A Survey 2025 found that 43 per cent of finance leaders had reduced or deferred finance headcount additions in 2025 due to AI productivity gains, compared with 18 per cent in 2023. The honest answer is that the productivity impact varies enormously by role and function, and blanket assumptions about AI productivity are not reliable. The right approach is to assess, function by function, where AI tools are genuinely reducing the time required for specific tasks, and to model headcount on the basis of output requirements rather than task volumes. Finance teams, for example, may be able to deliver the same management accounts and board pack with fewer junior finance staff if AI tooling automates reconciliations and variance analysis. This is a realistic assumption for 2026; it was not realistic for 2024.
Regulatory Compliance Costs as Explicit Budget Lines
A common budgeting failure at growth stage is to treat regulatory compliance as an implicit overhead rather than an explicit cost line. This leads to underestimating total compliance spend, making it invisible to the board, and failing to manage it actively.
For UK fintechs and growth companies in 2026, the following regulatory compliance costs should appear as named budget lines:
- FCA authorisation and supervision: Application fees, ongoing regulatory fees (based on annual income or other metric relevant to the firm type), FCA supervisory fees, and the cost of regulatory counsel and compliance consultants during any application process.
- Consumer Duty monitoring: The FCA's Consumer Duty, which came into full effect in July 2023, requires firms to monitor outcomes for retail customers on an ongoing basis. The cost of that monitoring, including any technology investment and staff time, should be a named line.
- DORA testing (where applicable): The Digital Operational Resilience Act applies to UK firms operating in or with EU entities from January 2025. Firms in scope should budget explicitly for ICT risk framework development, third-party risk management, and the required digital operational resilience testing programme.
- Data protection and AI regulation: The UK AI and Data Protection frameworks are evolving rapidly. Budget for external legal review of AI tooling and data handling practices, GDPR compliance audits, and any required remediation.
The Budget Calendar and Board Approval Process
A realistic 2026 budget calendar for a growth-stage company starting the process in October 2025 is as follows:
The board votes on the annual budget (the base case P&L, balance sheet and cash flow) and notes the downside scenario and the specific triggers that would prompt a formal review. The board does not approve the headcount plan in granular detail: that is an operating matter for management. What the board approves is the total headcount cost by function and the total non-headcount cost by category, embedded in the P&L.
Key Takeaways
- Driver-based budgeting is the most appropriate method for growth-stage companies: it makes assumptions explicit, enables rapid scenario modelling, and links operational plans directly to financial outcomes.
- Run an annual budget (for governance and investor reporting) alongside a quarterly 4+8 rolling forecast (for operational decision-making). The rolling forecast is not a budget revision.
- Build three explicitly distinct scenarios with stated assumptions. The value is not prediction but advance decision-making: knowing which levers to pull before the downside materialises.
- The October 2024 NIC changes must be explicitly included in 2026 headcount budgets. The employer NIC threshold fell to £5,000 and the rate increased to 15 per cent: model the incremental cost for your specific headcount profile.
- AI productivity gains are real but role-specific. Assess function by function rather than applying blanket assumptions. Finance teams can realistically deliver more output with fewer junior staff in 2026 compared to 2024.
- Regulatory compliance costs should be named budget lines: FCA fees, Consumer Duty monitoring, DORA testing, and data protection compliance are all material and must be managed actively.
- The board votes on the total budget (P&L, balance sheet, cash flow); operational detail is a management matter. The budget presentation to the board should take no more than 30 minutes and should focus on the key assumptions, not the line-by-line detail.
